The Worldwide Progress of SDGs. Depicting the Yearly Hot Topics, using Language Processing

The Sustainable Development Goals have become indispensable for the international agenda since their official launch in 2015. Corroborated, the publication of the yearly SDG reports represents a candid focal point for the gained progress, as long as a clear depiction of the global challenges, resources, and knowledge. Thus, in the 2016-2023 period, eight annual reports have been published, containing more than 500 content pages. Starting from those, we aim to design an explorative analysis framework, through the computation of several language processing techniques from the family of text mining, and sentiment analysis. Our objective will be to identify the top-level hot topics of every year, while also pinpointing certain temporal dynamics in terms of new or recurrent subjects, international threats, or social-economic discourse elements. Deep-diving into the methodological instruments, the analysis will be based on a mix of techniques, starting from the illustration of the most common terms and collocations. Afterward, we will apply the Term Frequency-Inverse Document Frequency algorithm to highlight the importance of several words relative to the entire collection of documents, with a focus on the evolutive dimension. Going forward, we will evaluate the sentiment scores for each report, to assess whether in certain moments the general tone of the conversation has switched to a more negative or positive valence. Throughout the applicative section, we will present the results through several graphical visualisations meant to better profile the connections between concepts. Finally, we consider the current approach to be valuable for understanding the general evolutions in terms of sustainable development, creating a summative and computational-based exploration of the progress generated worldwide in the last decade.